1. Transportation Engineering Mode Choice January 21, 2011

2. ANNOUNCEMENTS

3. No class until next Friday
Classroom will change to EN 3070 starting then

4. REVIEW

5. Trip generation: What are the total number of trips people make to and from each zone?
Trip distribution: What are the specific origins and destinations for this total number of trips?
Mode choice: How many people will choose to drive, walk, ride bicycles, use transit, etc.?
Route choice: What are the specific routes people will use for their trips?

6. So, we have more than one OD matrix.
AM PEAK
PM PEAK
OFFPEAK
(everything else)

7. Assumptions in this class:
Each work trip produces one journey from the home zone to the work zone during the AM peak, and one journey from work to home during the PM peak
AM Peak
PM Peak

8. Assumptions in this class:
Each shopping trip produces one journey from the home zone to the shopping zone during the off-peak, and one journey from shopping to home during the off-peak
Off-peak
Off-peak

10. Target we’re trying to get:
Example:
From the formula
Target we’re trying to get:

11. Previous adjustment factors
Example:
Previous adjustment factors 1 1 1 1
New factors:
12/13
16/19
48/68
116/92

12. Example: New table 12/13 16/19 48/68 116/92 New factors:

13. Example:
New table
12/13
16/19
(48/68) x (48/50)
(116/92) x (116/115)

14. MODE CHOICE

15. There are literally dozens of potential “modes” of travel:
Passenger vehicle
Drive alone
Carpool
Taxi
Walk
Bicycle
Public transit
Bus
Subway
Light rail
Streetcar
Commuter rail

16. There are literally dozens of potential “modes” of travel:
Airplane, helicopter, ferry, gondola, rickshaw, covered wagon, horse, etc. etc. etc.

17. DRIVE PERSONAL VEHICLE
To keep things simple, let’s just assume two modes:
DRIVE PERSONAL VEHICLE
vs.
BUS

18. To keep things simple, let’s just assume two modes:
vs.

19. What factors would affect this decision for you
What factors would affect this decision for you? Why might someone drive, or why might someone ride the bus?
Location (how close to a bus stop)
Convenience
Cost
Travel time
Parking
Health
Hard constraint
Privacy
Control

20. The fundamental idea is that people choose the option with higher “utility” for them.
Don’t have to find parking or worry about damage to car
Fun to drive, get there fast
Utility is a general economic term for “how happy you are with something” taking everything into account.

21. Different people and different trip purposes will lead to different utility.
(Road trip)

22. (The same framework can be applied to any other choice people make.)

23. (The same framework can be applied to any other choice people make.)
Great taste
Less filling

24. So, if we can estimate the utility each option provides to people, we can predict what choice they will make.
8,000 people will drive
3,000 people will take bus

25. We can define a utility function

26. Properties of mode m and calibrated coefficients
We can define a utility function
Utility of mode m
Properties of mode m and calibrated coefficients
Unobserved factors

27. Properties of mode m and calibrated coefficients
We usually focus on the observed utility
Utility of mode m
Properties of mode m and calibrated coefficients
and deal with the unobserved part later.

28. For instance
Income
Travel time
Cost

29. Now we can relate these to the census data, and our knowledge of the transportation system.
We need to generate total numbers of people driving or taking the bus. It’s easier to think in proportions Pcar and Pbus.

30. A few ground rules:

31. A few ground rules:
Pcar is increasing in Vcar
Pbus is increasing in Vbus

32. A few ground rules:
Pcar is increasing in Vcar
Pbus is increasing in Vbus
(because of unobserved part... at least some people will choose each option)

33. Perhaps the simplest function satisfying these rules is

34. Perhaps the simplest function satisfying these rules is
(typo in notes)

35. Proportion choosing car
This is the logistic curve: 1
Proportion choosing car
0.5
Vcar - Vbus

36. Check: ?

37. Check: ?
Pcar is increasing in Vcar
Pbus is increasing in Vbus

38. Check: ?

39. Continuing the example, here is the AM peak OD matrix:

40. With zone-to-zone travel times and costs by mode, we can use the equation to get mode split.
Travel time (min)

41. With zone-to-zone travel times and costs by mode, we can use the equation to get mode split.

42. We can then calculate the observed utility for each mode.
Census data

43. We can then calculate the observed utility for each mode.

44. Finally, we apply the logistic equation to get mode split.

45. Finally, we apply the logistic equation to get mode split.

46. If we multiply by the total number of trips going between each zone (dij), we get the total number of trips by mode: